last_chance_rockfandomcom-20200213-history
3D Scanning and its Uses
Introduction 3D scanners are primarily used to take information regarding the physical build of an object from a distance replacing the use of physical measurement tools such as calipers and CMMs, creating a digital model and study the original surface from it by taking 3-dimensional data of millions of points in order to reconstruct a surface in CAD (Computer-Aided Design) software. 3D scanning technology has made great leaps in the past few years, now being used in many different fields of work. The next generation of this technology requires scanners to be extremely accurate at great distances from the physical surface of the object. Future Mars rovers, require a 3d laser scanner that can collect great amounts of data, from large distances in order to make the task of navigating and mapping the terrain of mars efficient. An ideal model, with an extremely high level of accuracy is needed to set the standard for which the scanners shall compare themselves to. A 3D mesh and model is to be created of an artist’s representation of a rock from Mars using a 3D scanner from optimal distances in ordered to create a desired standard. 3D Laser Scanning ''' 3D laser scanners direct a laser beam in points closely spaced out, then measure the time of flight from the laser scanner to the physical object and back to the scanner. The location and position of the points is then established three dimensionally, thus creating “point clouds” which consist of thousand of point all representing the physical object dimensionally this information can be turned into a Computer-aided design (CAD) models that can be manipulated using CAD software (Arayici, 2007). This is done by turning “point clouds” into polygonal meshes by triangulation through the use of defined algorithms that calculate how the points should be connected. '''Use of 3D laser scanning technology 3D laser scanners are popular in their use for rapid and reverse prototyping, Arayici used one to scan a built environment and create a mesh then model and printout a 3D model of the building (2007). Terrestrial laser scanner have begun to be used more frequently and begun to shadow, other surveying techniques such as Electronic Distance Measurement (EDM), GPS and photogrammetric applications, for they are able to collect dense point data without the use of reflectors (Arayici, 2007). Merged the data clouds contain sufficient data to rid of any need for EDM interpolation, thus allowing for the optimal representation of the scanned surface. 3D laser scanner have also been used in efforts to create digital models of historical sights, an example would be the Digital Michelangelo project which became a challenge due the size of the object requiring several scans and a manner of joining them. Forensic scientists, physical anthropologist and conservators have found 3D scanners of great assistance in documenting, reconstructing, and analyzing objects and human remains. With spatial resolution matching or surpassing CT (Computed Tomography) and MRI (Magnetic Resonance Imaging) machines and since they use low energy non-hazardous light, they are capable of scanning skeletal remains (Sholts, Wamlander, Fores, Miller, & Phillip, 2010). Sholts, Wamlander, Fores, Miller, & Phillip designed a study to validate the use and repeatability of 3D laser scanner (2010). The results showed that the difference found between data of two different observers scanning craniums to find the surface area and value is minimal, falling below two standard deviations and with the difference between two scans made by the same person being less than .3% makes the use of 3d laser scanner reliable. The difference in data came from the manner of which the skulls where positioned and duration of scan. It can also come from errors in calibration and the method by which each observer did the scans. Anil, Tang, Akinci, & Huber, designed an experiment to determine the deviation in point cloud data by laser scanners finding various sources for which error could have occurred from (2013). Their deviation analysis was capable of determining of what physical errors amounted to a deviation in the data. This technology has also been employed in planet exploration such as on the Mars Rovers by NASA, exploration of dangerous terrain such as DEPTHX project by Carnegie Mellon University in Mexico and autonomous vehicles like Stanley in the DARPA Grand Challenge by Stanford University (Basaca-Preciado, et al., 2013). For it allows a rapid collectment of data to create a landscape for which the said vehicles calculate the ideal path to travel through. The scanners also play an important part in the exploration of the terrain. Point Cloud Interpretation A 3D laser scanner collects millions of raw points; these are converted into 3D meshes through the use of mathematical algorithms. Turning 3D scans into CAD models has many limitations. Once a scanner turns points into a polygonal mesh through triangulation, the mesh is not directly able to be turned into a 2D or 3D CAD model. The given software has to interpret the mesh into geometric shapes in order to set a uniform geometry rather than the millions of triangles in the mesh. The given 3D mesh also does not represent the physical object accurately at times, this being caused by “point cloud noise”, that is caused by physical intrusions such as a glossy surface sharp corners or curved features. When an object with complicated geometry is scanned the set algorithms are unable to interpret data very well creating noise and meshes and models that don’t match the physical structure of the object. Kalogerakis, Nowrouzezahrai, Simari, & Singh presented a method to create line of curvature on from point clouds with outliers (2009). Their method involved adapting to the determined data, trying to tell whether or not a curve is present then setting a uniform curve for the given points that fall near it to remove any intrusion in the curve. The complications that come with scanning sharp features is that, the existing methods are unable to reliably determine the normal, a perpendicular that is ninety degrees from the point, for points of sharp features since the area would have points belonging to different planes with their normals pointing in different directions (Wang, Feng, Delorme, & Engin, 2013). The normals of points are highly important; for they are required in common point based rendering techniques such as segmentation, smoothing, simplification, shape modeling, feature detection and extraction (Wang, Feng, Delorme, & Engin, 2013). Wang, Feng, Delorme, & Engin created method and algorithm for which normals can be better defined in order to create more accurate meshes and models of the physical object. Another challenge that comes with interpreting point clouds comes from when an object is so big it requires several scans, for the scans have to be aligned. One of the most common algorithms for pairwise rigid alignment was Iterative Closest Point (ICP) (Brown & Rusinkiewicz, 2007). ICP involves, starting off with a general alignment of then taking points from the models repeatedly finding the closest point in the opposite mesh. Global rigid body alignment aligns new scans to all of the previous vines. The error accumulates with successive scans and the more recent algorithms spread the error across (Brown & Rusinkiewicz, 2007). Turning 3D meshes to CAD models has proven to be difficult because CAD models are limited to the geometric primitives of planes, spheres, cones and cylinders, so when converting meshes the program is only finding the closest shape to the data collected. The method used by Beniere, Puech, Subsol, Gesquiere, & Le Breton, involves extracting geometric primitives, wire construction and model creation (2013). This simplification of steps allowed for a better monitoring of the data allowing for successful noise reduction and a better final CAD of the original object. Summary 3D scanners are very versatile tools with many advantages and possible uses. Though they work very well, they pose a possibility of inaccuracy that must be regulated and maintained to a minimum in the work they are used for. Several efforts have been made to adjust for these errors, through the use of different methods and algorithms. In order to continue this effort to make 3D scanning technology more accurate and reliable, one must create a baseline for which it should follow, thus reason for the creation of a 3D mesh and model of the said rock from Mars.